Sixteen children with moderate or severe bronchial asthma underwent a double-blind study receiving either ketotifen (K) or placebo (P) in order to verify the sparing effect of K on theophylline. The study lasted 18 weeks. Slow release theophylline was administered to all patients at a dose of 300 mg every 12 hours for the first six weeks of the study ("washout"). From the seventh week on, eight patients received K (1 mg, 2x/day) and the others P. The theophylline dose was maintained and reduced every 4 weeks to 200 mg and 100 mg. The initial clinical characteristics were identical in both groups. The serum levels of theophylline decreased below the ideal range (less than 10 mcg/ml) in groups K and P with a 100 mg dose of theophylline. However, in group K patients, fewer reported worsening of symptoms (number of days with cough and/or wheeze, altered daytime or nighttime peak expiratory flow rate and consumption of beta agonist) when compared to group P. Increased PC20 of methacholine only occurred in group K patients.